“AI Suicide”

Image from wikipedia https://en.wikipedia.org/wiki/File:Sophia_at_the_AI_for_Good_Global_Summit_2018_(27254369807)_(cropped).jpg of Sophia during the global AI for Good summit.
By Scott Hamilton
Last week I wrote about the halting of NASA’s quantum computing project, mainly due to shocking results that left the engineers questioning the safety of the technology. This week I came across another somewhat scary article on Artificial Intelligence (AI). There was another major breakthrough in AI in the last week, if you can call it a breakthrough.
This week in South Korea, a robotic civil servant employed by the Gumi City Council was found unresponsive at the bottom of a stairwell. The robot, nicknamed “Robot Supervisor,” worked a standard nine to five shift and even possessed its own civil service ID card. The robot was officially employed by the South Korean Gumi City Council.
A city official said, “It was officially a part of the city hall staff, one of us. It worked diligently.” The robot’s job was to deliver documents throughout the building, operating the elevators and delivering information to local residents. The robot has been in operation as part of the Gumi City Council staff since August 2023, and was developed by Bear Robotics. Bear Robotics is known for its full line of robotic waiters.
During the investigation into the robot’s demise, city officials were questioned regarding the robot’s schedule, if it were overworked or showing any signs of stress or unusual behavior. South Korea is known for its leading role in adopting robotic workers, with one industrial robot for every ten human employees on average in the country. However, Gumi City Council has announced that it will delay any plans to deploy additional robots in the wake of this incident.
Staff members present at the time of the robot’s demise said they noticed the robot circling as if it were looking for something, or avoiding an object that was not there. They had never before noticed such behavior in the robot and were worried that something was wrong. They considered the robot’s demise a suicide, mainly because none of the programming should have allowed the robot to enter the stairwell, much less fall to its demise. So why was the robot in the stairwell, and why did the sensors not prevent it from plunging to its untimely death?
I realize that there is a lot of speculation in calling this a suicide, because that would imply a free-will choice on the part of the robot to knowingly throw itself down the stairwell. I would have thought even a few years ago that this was impossible, because computers don’t think for themselves and only follow the instructions of their programmer. However, considering the recent advancements in AI, it is highly likely for a robot to do things outside of its preprogrammed operations… in this case, throw itself down the stairs.
If the investigation into the robot’s demise determines that there was, in fact, no prior damage to sensors, then we must assume that this was not a malfunction, but a decision by the AI programming to drive the robot over the edge. The simple fact that staff members referred to “Robot Supervisor” as one of the staff and referred to it by name makes it seem as if the AI within the robot gave it enough personality to be enjoyable to the staff members and community. The question left in one’s mind is if this AI became depressed enough to commit suicide, then is it equally possible that it could have become angry enough to commit murder?
I believe this is the question in the minds of the Gumi City Council as they consider the replacement of “Robot Supervisor.” Are the new AIs which are programmed to think for themselves, and even given the ability to express emotions, safe to interact with the general population? Are we on the brink of a war with AI? Are Isaac Asimov’s fictional predictions about robot behaviors becoming a reality? All these questions and more come to mind as the city of Gumi mourns the loss of their robotic co-worker.
Until next week, stay safe and learn something new.
Scott Hamilton is an Expert in Emerging Technologies at ATOS and can be reached with questions and comments via email to sh*******@**********rd.org or through his website at https://www.techshepherd.org.